Predicate AI Labs

About

Health signals have always told a story.
We listen.

Predicate AI Labs was founded on a simple insight: the human body signals deterioration long before it becomes clinically visible.

Our platform captures continuous voice and vital sign data, transforming it into time-aligned signals that reveal subtle physiological change.

Powered by our proprietary OpenDx™ engine, we detect patterns of health deterioration earlier—enabling clinicians to intervene before conditions escalate.

This is not another dashboard. It’s a new layer of clinical intelligence.

98%

sensitivity in identifying true sepsis
Prospective emergency department triage cohort (n=270) using voice-derived features in a low-resource setting

Clinical validation expanding across multiple geographies and care settings

// DATA INPUTS

Multimodal. Continuous. Patient-specific.

Voice Signals

Your voice as a physiological sensor. Detecting subtle changes in health through speech.

Vital Signals

Continuous physiological monitoring. HR, RR, Temp, SpO₂ from any connected device.

Time-Aligned Data

Signals synchronized across modalities. Capturing change as it happens.

Contextual Data

Clinical history, medications, labs. Used to enhance signal interpretation.

// CORE PROCESSING

AI

OpenDx™

Core Engine

1Data Fusion

2Pattern Mapping

3Risk Prediction

Platform

How It Works: From Signal to Early Insight

A real-time signal intelligence pipeline that transforms raw physiological data into early indicators of patient deterioration.

Multimodal inputs. Time-aligned. Continuously learned.

Signal Inputs

Vital Sign Streams

HR, HRV, RR, SpO₂, Temp — any wearable

Voice Features

Acoustic, prosodic & linguistic

Continuous Metabolic Monitoring

CGM data

Health History

Medications & labs

↓

Signal Processing

Noise Filtering

Adaptive artifact removal

Feature Extraction

Acoustic, linguistic, and physiological signals

Temporal Alignment

Multi-modal sync engine

↓

AI Engine

Baseline Modeling

Learns patient-specific physiological norms over time

Deviation Detection

Identifies subtle changes from baseline before clinical thresholds

Predictive Signal Mapping

Translates patterns into early indicators of deterioration

↓

Clinical Output

Real-Time Signals

Priority-ranked indicators of patient deterioration

Clinical Escalation Support

Risk-tiered guidance for timely intervention

Explainability Layer

Transparent signal attribution for clinician trust

Population Insights

System-wide visibility into patient risk trends

Where We Start

One problem.
One high-impact setting.

We are initially focused on one high-impact problem: identifying early signs of patient deterioration outside the hospital setting.

Starting with general practice / primary care, our platform enables clinicians to monitor patients between visits—when deterioration often goes unnoticed.

This creates a new standard of proactive care: continuous awareness without added clinical burden.

Team